How to implement a "function timeout" in Javascript - not just the 'setTimeout'

Question

How to implement a timeout in Javascript, not the window.timeout but something like session timeout or socket timeout - basically - a "function timeout"

A specified period of time that will be allowed to elapse in a system before a specified event is to take place, unless another specified event occurs first; in either case, the period is terminated when either event takes place.

Specifically, I want a javascript observing timer that will observe the execution time of a function and if reached or going more than a specified time then the observing timer will stop/notify the executing function.

Any help is greatly appreciated! Thanks a lot.

Answer 1

Here is my answer which essentially simplifies Martin's answer and is based upon the same tutorial.

Timeout wrapper for a promise:

const timeout = (prom, time) => { 
    const timeoutError = new Error(`execution time has exceeded the allowed time frame of ${time} ms`);
    let timer; // will receive the setTimeout defined from time 

    timeoutError.name = "TimeoutErr";

    return Promise.race([
        prom,
        new Promise((_r, rej) => timer = setTimeout(rej, time, timeoutError)) // returns the defined timeoutError in case of rejection

    ]).catch(err => { // handle errors that may occur during the promise race
        throw(err);

    }) .finally(() => clearTimeout(timer)); // clears timer 
}

A promise for testing purposes:

const fn = async (a) => { // resolves in 500 ms or throw an error if a == true
    if (a == true) throw new Error('test error');
    await new Promise((res) => setTimeout(res, 500));
    return "p2";
}

Now here is a test function:

async function test() {
    let result;

    try {  // finishes before the timeout
        result = await timeout(fn(), 1000); // timeouts in 1000 ms 
        console.log('• Returned Value :', result, '\n'); // result = p2
    
    } catch(err) {
        console.log('• Captured exception 0 : \n ', err, '\n');
    }

    try { // don't finish before the timeout

        result = await timeout(fn(), 100); // timeouts in 100 ms
        console.log(result); // not executed as the timeout error was triggered

    } catch (err) { 
        console.log('• Captured exception 1 : \n ', err, '\n');
    }

    try { // an error occured during fn execution time

        result = await timeout(fn(true), 100); // fn will throw an error
        console.log(result); // not executed as an error occured 

    } catch (err) {
        console.log('• Captured exception 2 : \n ', err, '\n');
    }
     
} 

that will produce this output:

    • Returned Value : p2 
    
    • Captured exception 1 : 
      TimeoutErr: execution time has exceeded the allowed time frame of 100 ms
        at C:\...\test-promise-race\test.js:33:34
        at async test (C:\...\test-promise-race\test.js:63:18)
    
    • Captured exception 2 : 
      Error: test error
        at fn (C:\...\test-promise-race\test.js:45:26)
        at test (C:\...\test-promise-race\test.js:72:32)

If you don't want to use try ... catch instructions in the test function you can alternatively replace the throw instructions in the catch part of the timeout promise wrapper by return.

By doing so the result variable will receive the error that is throwed otherwise.

You can then use this to detect if the result variable actually contains an error.

if (result instanceof Error) {
    // there was an error during execution
}
else {
    // result contains the value returned by fn
}

If you want to check if the error is relative to the defined timeout you will have to check the error.name value for "TimeoutErr".

Answer 2

My question has been marked as a duplicate of this one so I thought I'd answer it even though the original post is already nine years old.

It took me a while to wrap my head around what it means for Javascript to be single-threaded (and I'm still not sure I understood things 100%) but here's how I solved a similar use-case using Promises and a callback. It's mostly based on this tutorial.

First, we define a timeout function to wrap around Promises:

const timeout = (prom, time, exception) => {
    let timer;
    return Promise.race([
        prom,
        new Promise((_r, rej) => timer = setTimeout(rej, time, exception))
    ]).finally(() => clearTimeout(timer));
}

This is the promise I want to timeout:

const someLongRunningFunction = async () => {
  ...
  return ...;
}

Finally, I use it like this.

const TIMEOUT = 2000;
const timeoutError = Symbol();
var value = "some default value";
try {
  value = await timeout(someLongRunningFunction(), TIMEOUT, timeoutError);
}
catch(e) {
  if (e === timeoutError) {
    console.log("Timeout");
  }
  else {
    console.log("Error: " + e);
  }
}
finally {
  return callback(value);
}

This will call the callback function with the return value of someLongRunningFunction or a default value in case of a timeout. You can modify it to handle timeouts differently (e.g. throw an error).

Answer 3

I realize this is an old question/thread but perhaps this will be helpful to others.

Here's a generic callWithTimeout that you can await:

export function callWithTimeout(func, timeout) {
  return new Promise((resolve, reject) => {
    const timer = setTimeout(() => reject(new Error("timeout")), timeout)
    func().then(
      response => resolve(response),
      err => reject(new Error(err))
    ).finally(() => clearTimeout(timer))
  })
}

Tests/examples:

export function sleep(ms) {
  return new Promise(resolve => setTimeout(resolve, ms))
}

const func1 = async () => {
  // test: func completes in time
  await sleep(100)
}

const func2 = async () => {
  // test: func does not complete in time
  await sleep(300)
}

const func3 = async () => {
  // test: func throws exception before timeout
  await sleep(100)
  throw new Error("exception in func")
}

const func4 = async () => {
  // test: func would have thrown exception but timeout occurred first
  await sleep(300)
  throw new Error("exception in func")
}

Call with:

try {
  await callWithTimeout(func, 200)
  console.log("finished in time")
}
catch (err) {
  console.log(err.message)  // can be "timeout" or exception thrown by `func`
}

Answer 4

You could execute the code in a web worker. Then you are still able to handle timeout events while the code is running. As soon as the web worker finishes its job you can cancel the timeout. And as soon as the timeout happens you can terminate the web worker.

execWithTimeout(function() {
    if (Math.random() < 0.5) {
        for(;;) {}
    } else {
        return 12;
    }
}, 3000, function(err, result) {
    if (err) {
        console.log('Error: ' + err.message);
    } else {
        console.log('Result: ' + result);
    }
});

function execWithTimeout(code, timeout, callback) {
    var worker = new Worker('data:text/javascript;base64,' + btoa('self.postMessage((' + String(code) + '\n)());'));
    var id = setTimeout(function() {
        worker.terminate();
        callback(new Error('Timeout'));
    }, timeout);
    worker.addEventListener('error', function(e) {
        clearTimeout(id);
        callback(e);
    });
    worker.addEventListener('message', function(e) {
        clearTimeout(id);
        callback(null, e.data);
    });
}

Answer 5

I'm not entirely clear what you're asking, but I think that Javascript does not work the way you want so it cannot be done. For example, it cannot be done that a regular function call lasts either until the operation completes or a certain amount of time whichever comes first. That can be implemented outside of javascript and exposed through javascript (as is done with synchronous ajax calls), but can't be done in pure javascript with regular functions.

Unlike other languages, Javascript is single threaded so that while a function is executing a timer will never execute (except for web workers, but they are very, very limited in what they can do). The timer can only execute when the function finishes executing. Thus, you can't even share a progress variable between a synchronous function and a timer so there's no way for a timer to "check on" the progress of a function.

If your code was completely stand-alone (didn't access any of your global variables, didn't call your other functions and didn't access the DOM in anyway), then you could run it in a web-worker (available in newer browsers only) and use a timer in the main thread. When the web-worker code completes, it sends a message to the main thread with it's results. When the main thread receives that message, it stops the timer. If the timer fires before receiving the results, it can kill the web-worker. But, your code would have to live with the restrictions of web-workers.

Soemthing can also be done with asynchronous operations (because they work better with Javascript's single-threaded-ness) like this:

1. Start an asynchronous operation like an ajax call or the loading of an image.
2. Start a timer using setTimeout() for your timeout time.
3. If the timer fires before your asynchronous operation completes, then stop the asynchronous operation (using the APIs to cancel it).
4. If the asynchronous operation completes before the timer fires, then cancel the timer with clearTimeout() and proceed.

For example, here's how to put a timeout on the loading of an image:

function loadImage(url, maxTime, data, fnSuccess, fnFail) {
    var img = new Image();

    var timer = setTimeout(function() {
        timer = null;
        fnFail(data, url);
    }, maxTime);

    img.onLoad = function() {
        if (timer) {
            clearTimeout(timer);
            fnSuccess(data, img);
        }
    }

    img.onAbort = img.onError = function() {
        clearTimeout(timer);
        fnFail(data, url);
    }
    img.src = url;
}

Answer 6

Share a variable between the observing timer and the executing function.

Implement the observing timer with window.setTimeout or window.setInterval. When the observing timer executes, it sets an exit value to the shared variable.

The executing function constantly checks for the variable value.. and returns if the exit value is specified.

Answer 7

You can achieve this only using some hardcore tricks. Like for example if you know what kind of variable your function returns (note that EVERY js function returns something, default is undefined) you can try something like this: define variable

var x = null;

and run test in seperate "thread":

function test(){
    if (x || x == undefined)
        console.log("Cool, my function finished the job!");
    else
        console.log("Ehh, still far from finishing!");
}
setTimeout(test, 10000);

and finally run function:

x = myFunction(myArguments);

This only works if you know that your function either does not return any value (i.e. the returned value is undefined) or the value it returns is always "not false", i.e. is not converted to false statement (like 0, null, etc).